Hello Alex Docs

Creating Custom Voice Clones

Train reusable voice profiles from clean source audio. Read the Voices guide in Hello Alex documentation.

Overview

Voice choice changes trust and completion rates, so cloning is most useful when your team wants recognizable tone across launches.

Clean source audio is uploaded and processed into a reusable profile, which then appears anywhere voice can be selected in pathways and launchers.
Preview clips let you verify cadence and clarity before production assignment, reducing rework across active campaigns.

Good to know: cleaner recordings produce more stable pronunciation and fewer artifacts than long clips with room noise.

The sample stage exists to capture vocal identity accurately before the model is trained.

Run a send-call smoke test and compare intelligibility against your default voice.

You record or upload clear speech, name the profile, and submit it for processing so the generated clone can be reused in production surfaces.
A short preview confirms quality before assignment, which prevents mismatched tone from reaching live recipients.
Record vs Upload: In-browser recording or file upload.
Voice name: Label shown when picking the clone on pathways and launches.
Good to know: stable microphone distance and low background noise improve identity match more than longer clip duration.
Brand-matched receptionist voice
Founder voice for outbound campaigns

Assign the approved clone to a low-risk test pathway and validate with a real call.